Neuro-Fuzzy Segmentation of Document Images

P. Górecki (Italy, Poland), C. Castiello, and L. Caponetti (Italy)


Document Analysis, Neuro-Fuzzy Classification, Docu ment Image Segmentation, Feature Extraction.


The task of document image segmentation is to represent a digital image in a more interpretable form, recognising re gions containing text, background and graphics. This paper presents a peculiar strategy for document image segmenta tion, where a neuro-fuzzy approach is involved. Firstly, image is segmented into text, graphics or background dur ing a pixel level classification step. Successively, an analy sis performed over the obtained regions is devoted to re fine the initial segmentation results. A knowledge discov ery process is applied to automatically derive from sam ple data the fuzzy rule bases, responsible of the inference scheme presiding over the classification of image pixels and regions. The proposed method proves to be accurate and robust to page skew and noise.

Important Links:

Go Back